Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

CICERO: An AI agent that negotiates, persuades, and cooperates with people

ai.facebook.com

Post reply on HN
Print view View original

131–140 of 300 posts

Re: CICERO: An AI agent that negotiates, persuades, and cooperates with people

#131

Post by SteveDR »

gcanyon wrote: post #87

It seems like with this, Meta is doubling down on the Facebook platform's use as a dark influencer in politics. It's bad for humanity to enable advanced profiling and general targeting of demographics, but it's a whole other level if the communication itself can be targeted at the individual level. We're headed for a reality where I don't know if my friends' friends are real or not, and they each entice me with argum…

Do you think people are incapable of sticking to their beliefs when they’re faced with counterarguments?

Isn’t that kind of a pre-requisite for believing in democratic representation?

Re: CICERO: An AI agent that negotiates, persuades, and cooperates with people

#132

Post by yamtaddle »

bee_rider wrote: post #78
MichaelCollins wrote: post #41

Earlier quoted context omitted.

IMHO, Risk is one of the few board games worth playing specifically because it's so engaging. Most of my friends who are "into boardgames" tell me that Risk is a terrible game because it starts fights. But the games they want to play bore me to tears. (I can't even read the manual for Settlers of Catan without dying of boredom, but from what you say maybe I should give it another try.)

From what I’ve seen, the main complaint from board game enthusiasts about Risk is that the dice rolling is too random (I don’t really agree with this, the whole point of the game is to control your level of risk, there it is right in the name, so if your plan hinges on good rolls it is just a bad plan. But I digress). I think most people who are really into board games must have a general ability to separate in game…

Risk's much improved by goal-oriented or score-based gameplay variants, which tend to feature as the standard ruleset in newer versions and derivatives (Risk 2210, SPQRisiko, et c) for good reason. Conquer-the-world takes too long for a game that tends to eliminate half its players early on.

Re: CICERO: An AI agent that negotiates, persuades, and cooperates with people

#133

Post by daniel-cussen »

LanceH wrote: post #111
ghaff wrote: post #57

Earlier quoted context omitted.

>There is no real negotiation between players, other than an implicit, passing agreement when you don't provoke someone by attacking them. It depends how you play. I've seen people make treaties for all sorts of things for various durations, e.g. don't attack from Country X into Country Y and I won't attack across that border either.

Backstabbing in Risk or Catan is generally pretty obvious, though. It's more like, "Sorry, I need Australia." It's generally clear from the state of the game that it is headed there. In Diplomacy, it's a matter of choosing which of your friends to backstab. In a three way dynamic, it's basically two friends deciding that if the three of you crash in the mountains, they are eating you first.

Dude the World War Over Australia.

Some knock-off editions in Chile actually link...like it's generally called Perú, and I've never seen a map with Antartica, and besides Chile by it's extreme geography causes cartographers serious typographic problems. There was a map recently, a basically honest Mercator projection that replaced every country's territory with its name. It worked great, except for a mysterious country called CHILE CHILE CHILE CHILE CHILE. Broke the nomenclature completely.

Re: CICERO: An AI agent that negotiates, persuades, and cooperates with people

#135

Post by jfengel »

ThrowawayR2 wrote: post #30

I have to say that these advancements leave me unsure whether ML is much more advanced than I previously thought or human behavior is much less complex than I previously thought.

It's both, but I think the latter is the more notable part.

We like to pat ourselves on the back for behavior that computers can't do, but in a lot of ways those behaviors aren't really all that important. We pay attention to them because they're specific to us, but they're really just about humans interacting with each other rather than anything fundamental to the universe.

We're highly specialized for those things, and specialized by evolution which always takes a roundabout route to creating anything. So they're hard to reproduce precisely... but if you were creating an "intelligent species" from scratch you probably wouldn't have put them in there at all.

We still haven't really come close to understanding the weird, elliptical, Rube Goldberg mechanism by which brains produce consciousness. And that mechanism is neat for its ability to rapidly pick up certain categories of things -- albeit never logical, rigid categories of things. Anything logical or rigid can already be done by a computer a million times better.

Any time we carve out a sub-project it pretty quickly succumbs to a solution. Even off-the-wall stuff like "self driving cars" are really very good in 98% of cases after a couple of decades of concerted effort. We clutch our pearls about the 2% because we really hate it when people die... but that says as much about us as about the ML. If we were to replace ALL of the cars with AI, even with the 2022 top of the line, you'd probably get fewer net deaths.

So we still seem to be a long way from solving "human behavior". But it turns out that it may not actually be all that important, except as a bit of chauvinism defining ourselves as the most advanced thing in the universe.

Re: CICERO: An AI agent that negotiates, persuades, and cooperates with people

#136

Post by mistermann »

californiadreem wrote: post #69
EGreg wrote: post #64

Earlier quoted context omitted.

By the time GPT-4 botswarms infiltrate HN, Twitter, LinkedIn, Facebook and are weaponized to write articles that outshare the current media, it will be too late to stop them. They will amass more and bettee social capital, real capital, produce the vast majority of content and will be Among Us! My prediction is 2029

I'm afraid we're looking in the wrong places. See: http://txti.es/egregores

https://en.wikipedia.org/wiki/Pointing_and_calling

Re: CICERO: An AI agent that negotiates, persuades, and cooperates with people

#137

Post by KaoruAoiShiho »

mabbo wrote: post #28

I would love to see this kind of thing applied to an RPG. Randomly generate a city full of people. Make a few dozen of them the important NPCs. Give them situations and goals, problems they need to solve and potential ways to solve them. Certain NPC's goals are opposite others'. Then drop the player into that world and have the 'quests' the player is performing be generated based on the NPCs needing their help. Updat…

We already have this, it's just a matter of cost now.

Re: CICERO: An AI agent that negotiates, persuades, and cooperates with people

#138

Post by bee_rider »

vkou wrote: post #86
bee_rider wrote: post #9

Earlier quoted context omitted.

I have heard this about various games (Risk and Settlers of Catan mostly). I don’t get it. The point of these games is to be a bastard. If everyone goes into it with the understanding that there will be double crosses, how could there be hurt feelings? Monopoly is worse; it is just boring, I would dump my friends if they suggested monopoly not because I was hurt by their ruthless gameplay but for their terminally tas…

Settlers of Catan is in many respects not a very good game, because many of its mechanics are zero or negative-sum, with a heavy helping of RNG. This creates a lot of opportunities for kingmaking and spiteful plays from people who are not able to win, but are able to make sure that you lose. And the worst part is that you're often forced into these situations through no fault of your own. Diplomacy allows for all of…

I didn’t generally experience that. From what I saw, generally everyone gangs up on the person in the lead, if they are getting close to victory (this seems like a fair and obvious thing to do). So, part of the game is hiding how close you are until nobody can stop you. Hiding your strength is part of pretty much every strategy game with more than 2 players…

Is it a perfect game? No. But it is pretty good. And let’s not give Risk and Catan the Seinfeld treatment — sure there are better games… made in response to their perceived deficiencies!

Re: CICERO: An AI agent that negotiates, persuades, and cooperates with people

#139

Post by visarga »

bo1024 wrote: post #102
pesenti wrote: post #20

Paper: https://www.science.org/doi/10.1126/science.ade9097 Code: https://github.com/facebookresearch/diplomacy_cicero Site: https://ai.facebook.com/research/cicero/ Expert player vs. Cicero AI: https://www.youtube.com/watch?v=u5192bvUS7k RFP: https://ai.facebook.com/research/request-for-proposal/toward... The most interesting anecdote I heard from the team: "during the tournament dozens of human players never even su…

Hmm, I guess facebook doesn't have to go through IRB for human subject experiments, nor does Science require it, apparently.

That's nothing. Someone trained a GPT-J 6B model on 4chan and then let the model loose on the forums for a day. It took about 15k messages until people suspected something was off, and even that was because the country flags, Seychelles, a rare flag on 4chan, were a giveaway.

video: https://www.youtube.com/watch?v=efPrtcLdcdM

model: https://huggingface.co/ykilcher/gpt-4chan

Re: CICERO: An AI agent that negotiates, persuades, and cooperates with people

#140

Post by squeaky-clean »

yamtaddle wrote: post #112
bee_rider wrote: post #9

Earlier quoted context omitted.

I have heard this about various games (Risk and Settlers of Catan mostly). I don’t get it. The point of these games is to be a bastard. If everyone goes into it with the understanding that there will be double crosses, how could there be hurt feelings? Monopoly is worse; it is just boring, I would dump my friends if they suggested monopoly not because I was hurt by their ruthless gameplay but for their terminally tas…

You may be surprised that some people play board games assuming some kind of "fair play" or "play with fun in mind" rule is implicit. They may get very pissed off if you "ruin" a game by taking every single advantage you can according to the rules. [EDIT] Risk and Catan also have what are sometimes regarded as straight-up design flaws that can cause "zombie players" who stick around for most/all of the game but, very…

I've had the inverse happen to me, not playing optimally upsetting someone. Playing the DC superhero deck-building game at a card shop with some friends and a stranger asked to join, he already knew how to play.

Well I didn't stop another player from getting one of the final Power Ring cards and he flipped the F!&%$ out. "How could you not block her! She's guaranteed to win on her next turn now! You suck at this game!" Bro, I'm just trying to power up my batman. It's not a very competitive game.

Needless to say we asked him not to join in our next round, "that's fine, I don't want to play with a bunch of noobs."

Post reply on HN